| 摘要 |
PURPOSE: A method for distinguishing a letter area from a picture area in heterogenous mixed images is provided to distinguish a letter area from a picture area of heterogenous images, and to expand or contract heterogenous mixed images, thereby executing expansion/contraction of the mixed images without a big distortion. CONSTITUTION: A controller distinguishes a picture area from a letter area in heterogenous mixed images. The controller checks whether the distinguishing process of the letter area and the picture area is completed. If not, the controller covers a window to the heterogenous mixed images. When a 4-adjacent pixel grouping is completed, the controller retrieves the number of black pixel groups within the window. The controller retrieves a maximum black pixel group size. The controller retrieves whether the number of the black pixel groups is bigger than a threshold value '4' and smaller than a threshold value '3'. If so, the controller decides that a heterogenous mixed image area corresponding to the window is the picture area.
|
